Yield, nutrients uptake and lipid profile of the halophyte Salicornia ramosissima cultivated in two different integrated multi-trophic aquaculture systems (IMTA)

Aquaculture 583 (2024) 740547 Available online 14 January 2024 0044-8486/© 2024 The Authors. Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/bync-nd/4.0/). Yield, nutrients uptake and lipid profile of the halophyte Salicornia ramosissima cultivated in two different integrated multi-trophic aquaculture systems (IMTA) M. Castilla-Gavil´ an a , c , 1 , * , M. Mu˜ noz-Martínez a , 1 , E. Zuasti a , J. Canoura-Baldonado a , R. Mondo˜ nedo a , I. Hachero-Cruzado a , b a IFAPA - El Toru˜ no, Camino Tiro Pich´ on s/n, El Puerto de Santa María, C´ adiz, Spain b “Crecimiento Azul”, Centro IFAPA el Toru˜ no, Unidad Asociada al CSIC, El Puerto de Santa María, C´ adiz, Spain c Laboratorio de Biología Marina, Departamento de Zoología, Facultad de Biología, Universidad de Sevilla, Avda. Reina Mercedes 6, Sevilla, Spain ARTICLE INFO Keywords: Aquaculture wastewater valorization Nutrient film technique (NFT) Deep water culture (DWC) Nitrogen uptake efficiency (NUE) Nitrogen uptake rate (NUR) Fatty acids ABSTRACT The nutrient removal, biomass production and nutritional value of the halophyte Salicornia ramosissima were evaluated in the context of an Integrated Multi-Trophic Aquaculture (IMTA) system. The experiment consisted of an indoor fish RAS whose effluent was used to feed two different halophyte hydroponic systems: a Nutrient Film Technique (NFT) system or a Deep Water Culture (DWC) system. The highest yield of S. ramosissima was observed in the DWC (320 g m −2 day −1 ), this system showed significantly higher yields compared to the NFT. Both hydroponic systems were effective in removing nutrients from the fish RAS wastewater, contributing to the wastes valorization process. Maximal nutrient uptake efficiencies were: 82.8% and 68.6% NO 3 − in the DWC and the NFT, respectively; 72.6% PO 4 3− in the DWC and 82.9% in the NFT; 63.6% NO 2 − in the DWC vs 56.4% in the NFT; and 89.2% NH 4 + in the DWC vs 94.1% in the NFT. The lipid profile of the halophytes in the DWC was analysed. The total lipid content ranged from 2.99% to 3.83%. The fatty acids (FA) profile in the leaves was dominated by polyunsaturated fatty acids (about 66%), which were represented by α -linolenic acid (41%) and linoleic acid (24%), followed by saturated fatty acids (28%) with palmitic acid (22%) as the major component. A small component of monounsaturated fatty acids accounts for about 12% of total FA. Regarding lipid classes, pigments were the most abundant compounds followed by glycolipids, phospholipids, flavonoids and sterols. The study supports the potential of S. ramosissima to valorise aquaculture wastes and its high nutritional interest. The experimental systems minimized waste and maximized resources utilization. Such integrated approaches can contribute to the sustainability and resilience of aquaculture systems while providing additional economic opportunities through the production of high value halophyte products. 1. Introduction The aquaculture industry produces more than half of the fish and seafood for human consumption (FAO, 2022) and plays a fundamental role in meeting to the growing demand for protein and essential fatty acids of the population. To meet this high demand in an environmentally, socially and economically sustainable way, several innovative production techniques and policies have been improved in the last years (EU missions “Restore our Ocean and Waters” or “Green Deal”, the Atlantic Action Plan or the Blue Growth strategy; European Commission, 2021, 2020, 2019, 2013). More recently, the United Nations Sustainable Development Goals aim to move towards a circular economy through Goal 12. This model implies a change in the consumption and production paradigm, where resources become products, products become waste and waste becomes resources (United Nations, 2023). The present study falls within this framework of waste valorisation. Recirculating Aquaculture Systems (RAS) are well known to be highly efficient systems, allowing intensive land-based production at * Corresponding author at: Laboratorio de Biología Marina, Departamento de Zoología, Facultad de Biología, Universidad de Sevilla, Avda. Reina Mercedes 6, Sevilla, Spain. E-mail addresses: [email protected], [email protected] (M. Castilla-Gavil´ an). 1 These authors contributed equally to the research. Contents lists available at ScienceDirect Aquaculture journal homepage: www.elsevier.com/locate/aquaculture https://doi.org/10.1016/j.aquaculture.2024.740547 Received 7 September 2023; Received in revised form 15 December 2023; Accepted 3 January 2024 Aquaculture 583 (2024) 740547 2 high densities with a reduced water consumption and pollutant release (Ahmed and Turchini, 2021). In RAS, the water filtration processes significantly reduce the presence in the effluent of ammonium and nitrite (Badiola et al., 2012; Martins et al., 2010), highly toxic compounds for the cultured species and the environment (Rodrigues et al., 2007; Sampaio et al., 2002). Despite of this great management of the wastewater, nitrate and phosphate can also cumulate in the RAS effluents (Beyer et al., 2021). Although these inorganic compounds are less harmful than nitrite and ammonium (Camargo et al., 2005), they are the limiting resources preventing excess of plants growth in marine environments and, at high concentrations, they can cause eutrophication and algal blooms (Farmer, 2018). Furthermore, the degradation of the suspended solids in the wastewater can lead to high biological oxygen demand and the production of ammonium due to mineralisation (Buhmann and Papenbrock, 2013a). Aware that the solutions to reduce wastewater are usually costly for farmers, RAS are good candidates to be combined with Integrated Multi- Trophic Aquaculture (IMTA; Li et al., 2019; Sri-uam et al., 2016). IMTA systems consist of two or more functional groups that are trophically linked by nutrient flows (Dunbar et al., 2020), allowing groups of the lower trophic levels to feed on the wastes produced by those on the upper levels. Integrating the cultivation of at least a photosynthetic species into the RAS production cycle will make it possible to reduce inorganic pollutants and to produce a new extractive species of high market value. In marine IMTA, the use of halophytes as highly salt-tolerant species has already been highlighted (Cust´ odio et al., 2017; Senff et al., 2020; Shpigel et al., 2013; Waller et al., 2015). Halophytes have been harvested and consumed as a source of food and for their medicinal Fig. 1. A) Experimental design. Green lines represent water flow from the indoor RAS to the S. ramosissima culture systems under a greenhouse. Red points represent water inflow sampling points. Blue points represent water outflow sampling points. Yellow points represent plants sampling points for lipid analysis. B) Both systems (DWC on the left, NFT on the right) under a greenhouse. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.) Table 1 Water flow rates during the trial in both systems. June July August September NFT (L day −1 ) 375 400 400 400 DWC (L day −1 ) 375 400 600 600 NFT =nutrient film technique; DWC =deep water culture. M. Castilla-Gavil´ an et al. Aquaculture 583 (2024) 740547 3 applications since ancient times (Ventura and Sagi, 2013), and their cultivation has been improved in recent decades (de Vos et al., 2010; Debez et al., 2010; Herppich et al., 2012; Ventura et al., 2011a; Wilson et al., 2000) due to the increasing arid and semi-arid areas, soil salinization and freshwater scarcity (Koyro et al., 2011; Rozema and Flowers, 2008). The cultivation of halophytes is of great economic interest due to their potential uses in gourmet cuisine and pharmacology (Buhmann and Papenbrock, 2013b; Casta˜ neda-Loaiza et al., 2020). Indeed, in addition to their high-valued organoleptic properties, these species have high contents of proteins (Jer´ onimo et al., 2021), n-3 and n-6 polyunsaturated fatty acids (Isca et al., 2014), minerals, vitamins (Barreira et al., 2017) and polyphenolic compounds associated with antioxidant activity and health benefits (Ksouri et al., 2012). These compounds are generally stimulated in response to biotic/abiotic stresses (Grigore and Oprica, 2015; Navarro et al., 2006) and their concentration could be influenced by cultivation conditions. In the geographical area of the present study, Cadiz Bay (36.43875, −6.17713), several halophyte species grow naturally in marsh ecosystems (Curado Ferrera, 2012; Curcio et al., 2023). The species Salicornia ramosissima J. Woods, commonly known as “sea asparagus”, is particularly appreciated and is used in both popular recipes and haute cuisine (Feng et al., 2013). Salicornia ramosissima is an annual plant that tolerates high salinities (Barreira et al., 2017), with a nutritional profile suitable for human consumption and a potential for the valorization of Fig. 2. A) Detail of one NFT bench. B) Detail of DWC tank lines. Table 2 Water parameters in both systems throughout the trial (mean ±SD). Salinity Dissolved oxygen (ppm) pH Temperature (◦C) NFT 45.5 ±5.2 6.8 ±0.3 7.6 ±0.3 24.3 ±1.9 DWC 44.4 ±1.6 4.3 ±1.7 7.3 ±0.3 23.7 ±1.5 NFT =nutrient film technique; DWC =deep water culture. Fig. 3. Salicornia ramosissima from germination to transplantation: A) Germinated seedlings; B) Four-months aged seedlings; C) Rock wool cubes; D) Seedlings batches in their pots; E) Five-months aged seedlings placed in the system. M. Castilla-Gavil´ an et al. Aquaculture 583 (2024) 740547 4 the aquaculture effluents (Jer´ onimo et al., 2021). Hydroponic cultivation of halophytes has previously been underlined as promoting a high efficiency of nutrient removal from aquaculture effluents compared to conventional substrate-based cultures (Buhmann and Papenbrock, 2013a). Marques et al. (2021) highlighted a similar nutrient profile in halophytes cultivated with marine wastewater or wild plants. Thus, hydroponic cultivation of S. ramosissima by valorising aquaculture effluents would allow, besides their remediation, a high quality and a more sustainable production in areas where water is scarce (Ventura and Sagi, 2013). Different hydroponic techniques are usually used depending on the need for biofiltering, the nutritional requirements of the plants or the availability of land by the farmers (Schmautz et al., 2016). Few studies have compared different hydroponic techniques in an aquaculture context (Goda et al., 2015; Lennard and Leonard, 2006; Schmautz et al., 2016) and, to our knowledge, there are no studies comparing these techniques in the cultivation of halophytes. The present work aimed to implement an IMTA system with Solea senegalensis and Dicentrarchus labrax RAS production and hydroponically cultivated S. ramosissima for the valorization of the RAS effluent, promoting the sustainable culture of local products in a circular blue economy model. Our first objective was to compare two different hydroponic systems to determine which one allowed a higher 1) efficiency of S. ramosissima in removing nutrients from the fish RAS wastewater and 2) plant productivity and growth rates. The second objective was to evaluate the nutritional profile of S. ramosissima grown under these conditions. For this purpose, halophytes were grown with the RAS effluent in a Nutrient Film Technique (NFT) system or in a Deep Water Culture (DWC) system. In the last decades, DWC cultivation, in which plants are fixed in polystyrene rafts floating on the surface of the water (Forchino et al., 2017), has been widely used because it provides a large space for plant roots and an efficient nutrient delivery to the plants (Rakocy, 2012). On the other hand, NFT systems consist of channels with a flowing thin film of water in which plant roots are partially submerged. This technique is said to offer advantages over DWC such as simpler, lighter, less bulky and more economical construction by allowing higher densities (Lennard and Leonard, 2006). In our research, fish and plant productivity, growth rates, nutrient system removals and uptake efficiencies were monitored during a four-month experiment. In addition, the lipid profile of the halophytes was carried out to determine differences in their composition throughout their life cycle or their position in the water mass of the system. 2. Materials and methods 2.1. Experimental design The experiment lasted 16 weeks from June to September 2022. The IMTA system consisted of an indoor fish RAS whose effluent was used to feed two different halophyte hydroponic systems: NFT or DWC (Fig. 1). The RAS consisted of four units of four 350 L fish tanks equipped with a cooler, mechanical filter (sand filter), skimmer, ultraviolet lights and biofilter. Discarded water from the RAS was pumped twice a day (8 am and 1 pm) into separated 550 L distribution tanks equipped with particles retention compartments (Fig. 1). From the third month and taking in to account that DWC systems need greater water volumes than other hydroponic systems (Surnar et al., 2015), effluents from three random RAS units were pumped to the rafts and the effluent from the fourth unit was pumped to the NFT. Table 1 shows the daily flow rates over the fourmonth period in each system. These flow rates were discontinuous, water took only a few hours to get into the systems from the distribution tanks and then remained quiet. Water from the distribution tanks was then pumped into the different systems. NFT system consisted of three-replicated benches of NFT hydroponic channels; each bench was 2 m in length, 60 mm in width, and 50 mm in depth, with 6 channels per bench (total surface area per bench of 2.1 m 2 , with a total production area of 6.3 m 2 ). The total number of plant growth holes in this system was 558 (31 holes per channel), and the plant spacing was 130 mm based on commercial recommendations (Figs. 1; 2A). The DWC system consisted of two-replicated lines; each line with four 250 L-1 m 2 tanks connected by water flow and equipped with floating polystyrene rafts. Total surface area per line was 4 m 2 with a total production area of 8 m 2 . The total number of plant growth holes in this system was 288 (36 holes per raft) and the plant spacing was 150 mm (Figs. 1; 2B). All this experimental array was placed in a greenhouse, under natural photoperiod and an air temperature varying between 25 and 40 ◦C throughout the trial. Water parameters (salinity, dissolved oxygen, pH and temperature) were checked weekly (Table 2). 2.2. Reared fishes, sampling and biometrics assay At the beginning of the experiment, each RAS unit contained 70 kg (50 Kg m −3 ) of the Senegalese sole S. senegalensis. Four weeks later, the sole biomass reached the maximum capacity of the system. For this reason, and to ensure the welfare of the fish in the tanks, they were sampled, biometric measurements were performed (weight, length) and they were replaced by 11 kg (10 Kg m −3 ) per RAS unit of sea bass D. labrax. Sea bass were sampled and measured at the end of the trial. Feeding rates were in average 540 g day −1 per RAS unit for soles and 121 g day −1 for sea bass. Foods were formulated in a protein:lipid ratio of 51:15 for sole and 52:17 for sea bass. Specific Growth Rate (SGR) and Food Conversion Rate (FCR) were then calculate following García- Meil´ an et al. (2013). Table 3 Salicornia ramosissima biometrics in both systems at the beginning of the trial (mean ±SD). Plant length (cm) Root length (cm) Leaf area (cm 2 ) Total weight (g) NFT 11 ±1.8 14 ±4.9 47 ±20.7 118 ±46.2 DWC 16 ±2.0 20 ±3.6 141 ±34.0 329 ±118.8 NFT =nutrient film technique; DWC =deep water culture. Table 4 Water flow characterization during the trial in both systems. June July August September NFT ( μ mol L −1 ) Nitrate 319.1 ± 86.5 696.9 ± 469.7 1148.5 ± 525.4 1140.9 ± 107.9 Nitrite 39.4 ±26.4 11.9 ±12.3 17.9 ±8.1 41.9 ±33.3 Ammonium 34.8 ±25.9 11.9 ±7.9 20.4 ±4.3 40.8 ±18.6 Phosphate 19.9 ±10.1 65.0 ±15.8 43.7 ±33.0 84.4 ±69.2 DWC ( μ mol L −1 ) Nitrate 347.9 ± 65.6 614.1 ±72.8 1143.5 ± 212.9 1281.6 ± 327.2 Nitrite 38.9 ±25.7 12.9 ±6.8 20.5 ±5.3 28.7 ±6.7 Ammonium 21.5 ±22.8 13.2 ±5.8 20.1 ±1.4 28.7 ±6.5 Phosphate 19.9 ±10.1 65.0 ±15.8 42.8 ±33.3 66.9 ±47.9 NFT =nutrient film technique; DWC =deep water culture. Data are mean ±SD. Table 5 Mortality of S. ramosissima cultured in NFT or DWC systems during the trial. NFT DWC June 0% 0% July 2.63% 0% August 0.88% 2.09% September 41.24% 2.08% NFT =nutrient film technique; DWC =deep water culture. M. Castilla-Gavil´ an et al. Aquaculture 583 (2024) 740547 5 2.3. Halophytes cultivation, sampling and biometrics assay Salicornia ramosissima seedlings were provided by Hal´ ofitas Onuba S. L. (Isla Cristina, Spain). Four-month-old S. ramosissima seedlings were transplanted in April 2022. They were removed from their seedbeds and their roots were carefully rinsed with seawater in order to remove soil residues. Batches of 4 to 6 seedlings were then transplanted into 5 cm diameter plastic pots with rock wool as a substrate. Pots were randomly placed in the 846 holes prior to the experiment (Fig. 3). In June 2022 plants were labelled and a sample of 276 plants was randomly selected in order to perform biometrics and biochemical analyses at the beginning of the experiment (June) and every month (July, August and September): 180 plants from the NFT (10 plants per channel) and 96 plants from the DWC (12 plants per tank). All plants were pruned (only new edible shoots were cut) at every sampling period and biometrics (total weight) and photographs were performed on the sampled plants before and after being pruned. Photographs were used to estimate the plants green leafs area (%) by a colour threshold image segmentation with the Fiji ImageJ software (Schindelin et al., 2015; Sibiya and Sumbwanyambe, 2019). Yield was then calculated for both systems units (one channel for the NFT system and one tank for the DWC system) as fresh biomass accumulation per m 2 and day (Ventura et al., 2010). As a growth gradient was observed between the DWC plants placed in the tanks close to the water inflow and those close to the water outflow, it was decided to process and present the data grouping the plants by their position in the systems (inflow, middle, outflow; Fig. 1). Halophytes at the beginning of the experiment (June) presented the following biometric characteristics (Table 3). 2.4. Water chemistry Triplicated water samples were collected every week (16 sampling periods in total) at all inflow and outflow points of both systems (Fig. 1A). Water samples were filtered on GF/C fibreglass filters that were treated following Castilla-Gavil´ an et al. (2023) in order to determine total suspended solids in the water and the organic matter content of this solids. Water was then analysed: dissolved nitrate and nitrite concentrations were sequentially quantified by the reduction of NO 3 − to NO 2 − with vanadium (VCl 3 ) following the method of García-Robledo et al. (2014). Ammonium and phosphate concentrations were determined by the indophenol-blue method (Aminot et al., 1997) and the Murphy-Riley reaction method (Murphy and Riley, 1962), respectively. Fig. 4. Salicornia ramosissima life cycle under cultivation in NFT or DWC. A) Vegetative stage in DWC (July–August). B) Vegetative stage in NFT system (July–August). C) Flowering stage (August–September). D) Leaf senescence (September, end of the trial). E) Death (October, one month after the end of the trial). M. Castilla-Gavil´ an et al. Aquaculture 583 (2024) 740547 6 Results of the 16-week analyses were averaged over groups of four weeks to facilitate data processing and presentation. The groups were named by months (June, July, August and September). Nutrients uptake efficiency (NUE) and nutrients uptake rates (NUR) were calculated following Chu and Brown (2022) and Webb et al. (2013), respectively, as follows: NUE (%) = 100 xCi −Co Ci being Ci the nutrients (N or P) concentration in the water inflow (g L −1 ) and Co the nutrients concentration in the water outflow (g L −1 ). NUR (mmol m−2day−1)=Q x(Ci −Co S) being Q the water flow rate (L day −1 ) and S the surface in m 2 . 2.5. Halophytes lipid profile Considering mortality, yield and nutrients uptake results in the two different systems, biochemical analyses were only performed on the edible shoots pruned every month from plants in the DWC system placed in the inflow and outflow positions (Fig. 1). For each sampling period, fresh samples were pooled by position (in, out) and ashes, moisture, total lipids, fatty acid profile and lipid classes profile were analysed. Total lipids (TL), fatty acids (FA) and lipid classes (LC) profile methods were those previously described by García-Garrido et al. (2010) and Hachero-Cruzado et al. (2014). Briefly, TL were extracted from plants powder following the methodology of Folch et al. (1957) with slight modification by Christie and Han (2012). To extract FA, total lipid extracts were subjected to acid-catalysed transmethylation. Fatty acids methyl esters (FAME) were separated and quantified by gas chromatography and the identified FA were compared to the authentic standards (FAME Mix C37) and well-characterized fish oil (Menhaden Oil, SUPELCO, USA) (García-Garrido et al., 2010). Lipid classes were separated using one-dimensional double development high-performance thin layer chromatography (HPTLC). The first development was separated through Olsen and Henderson (1989) method with a following polar solvents mixture methyl acetate/isopropanol/chloroform/methanol/0.25% (w/v) KCl (25:25:25:10:9). Lipid classes were quantified by charring with a copper acetate reagent followed by calibrated scanning densitometry using a CAMAG TLC Scanner 3. Individual lipid spots were visualized through a UV fluorescence at 366 nm. Finally, authentic standards (Sigma-Aldrich) were used to determine the phospholipids, glycolipids and neutral lipids. 2.6. Data analyses Mean and standard deviation (SD) were calculated for all data. Statistical analyses and plots were performed using RStudio software. A three-way mixed repeated measures ANOVA was performed to Fig. 5. Yield (g m −2 day −1 ) of S. ramosissima cultivated in NFT or DWC systems during the trial. Error bars represent standard deviation. In =inflow position; mid =middle position; out =outflow position. NFT =nutrient film technique; DWC =deep water culture. Table 6 Results of three-way mixed repeated measures ANOVA for yield.* , ** Source of variation DFn DFd F value P value System (S) 1 221 422.31 0.000*** Position (P) 2 221 0.06 0.942 Time (t) 1.94 429.84 125.46 0.000*** S:P 2 221 0.47 0.628 S:t 1.94 429.84 32.74 0.000*** P:t 3.89 429.84 21.80 0.000*** S:P:t 3.89 429.84 19.68 0.000*** Levene Test jul: 5 273 10.3 0.000 aug: 5 260 15.7 sept: 5 243 33 * P <0.05. ** P <0.01. *** P <0.001. M. Castilla-Gavil´ an et al. Aquaculture 583 (2024) 740547 7 evaluate the effects of the system (two levels: “NFT” and “DWC”), the position in the system (three levels: “in”, “mid” and “out”) and the time (three levels: “July”, “August” and “September”) on the plants yield. System and position were the two between-subjects factors and time was the within-subjects factor. Where ANOVA indicated significant interactions (p <0.05), two-way interactions and simple simple main effects tests were carried out with a Bonferroni adjustment leading to statistical significance being accepted at the p <0.025 level. All simple simple pairwise comparisons were run between the different factors and a Bonferroni adjustment was also applied. Two-way ANOVAs were run for testing effects of the system (two levels: “NFT” and “DWC”) and the time (four levels: “June”, “July”, “August” and “September”) on the halophytes NUE and NUR. Where ANOVA indicated a significant difference (p <0.05) for a given factor, the source of the difference was identified using a Tukey test. Principal component analysis (PCA) was conducted to FA and LC matrixes for the ordination of samples. Additionally, permutation tests for multivariate analysis of similarity (PERMANOVA) were conducted to explore significant differences in lipid composition among time (four levels: “June”, “July”, “August” and “September”) and position (two levels: “in” and “out”). Two-way ANOVAs were also performed for individuals FA and LC with factors time and position. Prior to ANOVAs, the normality was tested with Shapiro-Wilk test (p >0.05) and the homogeneity of variances was tested with Levene test (p >0.05). For the repeated measures ANOVA, Mauchly test of sphericity and sphericity corrections by adjustments to the degrees of freedom were run when needed. Since proportions were compared, an arcsine transformation was carried out (Sokal and Rohlf, 1995). If variances remained heterogeneous even after data transformation, untransformed data were still analysed, as ANOVA is a robust statistical test and is relatively unaffected by the heterogeneity of variances (Underwood, 1997). In such cases, the level of significance was reduced to <0.01 to avoid type I error. 3. Results 3.1. Fish growth and RAS effluent characterization Senegalese soles and sea bass survival was observed to be higher than 95%. Soles biomass reached the maximum capacity of the system (79 kg per unit, 56 Kg m −3 ) in four weeks. For this reason, they were harvested and replaced by sea bass. At the end of the trial (i.e. 12 weeks later), sea bass biomass reached 43 kg (40 Kg m −3 ) per unit. Specific growth rates were 0.35% day −1 for soles and 1.74% day −1 for sea bass. Food conversion rates were 1.74 for soles and 1.67 for sea bass. The RAS effluent characterization over the four-month trial in each system is presented in the Table 4. 3.2. Halophytes survival and yield Mortality was observed mostly in the NFT system by the end of the trial (between August and September; Table 5). Salicornia ramosissima grew in both systems and all its life cycle stages were observed during the trial (Fig. 4). Maximal productivity was observed during the vegetative stage (July–August), when an average of 2.2 Kg m −2 and 3.5 Kg m −2 were harvested per month in the NFT and the DWC system respectively. Roots in the NFT at the end of the experiment were 19.6 ±4.6 cm length and 4.5 ±0.5 width. Roots in the DWC were Fig. 6. Salicornia ramosissima nitrate, phosphate, nitrite and ammonium nutrient uptake rates (mmol m −2 day −1 ) under cultivation in the NFT system (black bars) or in the DWC system (grey bars) during the four-month trial. Error bars represent standard deviation. NFT =nutrient film technique; DWC =deep water culture. M. Castilla-Gavil´ an et al. Aquaculture 583 (2024) 740547 8 28.7 ±6.8 cm length and 4.6 ±0.5 width. Yield is represented in the Fig. 5 and the results of the three-ways mixed ANOVA are shown in the Table 6. The DWC system showed a significantly higher yield than the NFT (p <0.001). In the NFT, yield significantly decreased over time (p <0.001) at all three positions (Table 6). Significant three-way interactions were observed between factors (p <0.001; Table 6): there was a growth gradient in the DWC system causing a delay in the development of plants that were further away from the nutrient inflow (yield in July was significantly lower in the outflow tanks (Df =2, F =20.6, p <0.001). Contrarily, in September, yield was significantly higher in these tanks (Df =2, F = 11.9, p <0.001). 3.3. Suspended solids No particles retention by plants or decantation was observed in any system throughout the trial. Total suspended solids in the water varied between 492 and 374 mg L −1 in the NFT system and between 475 and 374 mg L −1 in the DWC system, without significant differences between them (Df =1, F =3.4775, p =0.0641) nor between inflow and outflow water (Df =1, F =2.7617, p =0.0986). On average, 439 mg L −1 of suspended solids got into the NFT system, of which 24% was organic matter (OM). Concerning DWC, an average of 435 mg L −1 got into the system (25% OM). No differences were found between OM content of the suspended solids in both systems (Df =1, F =1.4994, p =0.2225) nor between water inflow and outflow (Df =1, F =2.9605, p =0.0872). 3.4. Nutrient uptake efficiency and nutrient uptake rate In general, S. ramosissima presented a high capacity of nutrient absorption for all nutrients and sampling periods. Maximal nitrate NUE was observed in the DWC system during the halophytes vegetative stage (82.8%). Significant differences were only observed between DWC and NFT (68.6%) nitrate NUE in June (Df =3, F =6.7997, p <0.05). For phosphate NUE, maximal values were also observed during vegetative stage of plants, without significant differences between systems (82.9% for NFT and 72.6% for DWC; Df =1, F =3.0462, p =0.0851). Nitrite and ammonium were also efficiently absorbed: maximal values for nitrite were 56.4% in the NFT system and 63.6% in the DWC, without significant differences (Df =1, F =0.0258, p =0.8728). Maximal ammonium NUE was significantly higher in the NFT system (94.1%) than in the DWC (89.2%) (Df =1, F =5.6978, p <0.05). Nutrient uptake rates observed in the four sampling periods are represented in Fig. 6. In general, uptake was significantly higher in the DWC system for all nutrients (Table 7). Nitrate uptake was significantly higher in August, which coincides with the higher amount of nitrate present in the effluent from this time (Table 4). The same trend was observed for phosphate uptake from the second month, although not statistically supported (Table 7). For nitrite and ammonium, halophytes also show uptakes related to the nutrients amount present in the water flow (Table 4). In both cases, uptakes were lower during July and August (Table 7). 3.5. Lipids profile characterization Halophytes total lipids content in the DWC system was in average between 2.99% and 3.83% (Fig. 7). During the first month of the trial, a significantly higher lipid content was observed in the plants placed at the inflow position tanks (Table 8) showing again a gradient between plants in the inflow or outflow positions. The most abundant FA were: the polyunsaturated (PUFA) 18:3n3 (ALA) and 18:2n6 (LA), and the saturated (SFA) 16:0 (PA) (Table 9). ALA and LA presented contrary tendencies. They remained stable over the experiment until September, when ALA significantly decreased and LA significantly increased (two-way ANOVAs; Table 10). Halophytes in the inflow position presented significantly higher proportion of PA than those in the outflow position (two-way ANOVA; Table 10). PERMANOVA reflected significant differences in fatty acid composition between sample periods (i.e. June, July, August and September; F =18.207, p =0.001) and positions (F =7.949, p =0.002). Pairwise comparisons revealed significant differences between September samples and June, July and August samples (p <0.010), whereas no differences were found in pairwise comparisons of positions (p =0.068). However, the test detected interactions between factors (F =2.4081, p =0.048) and finally differences between positions were detected within July samples (F =2.874, p =0.027) and August samples (F =2.502, p = 0.033). These differences were also supported by the PCA analysis. The first principal component explained 86.2% of the total variance and the second principal component explained 6.4%. Principal component 1 (PC1) correlated positively with the fatty acid 18:3n3 (ALA) (r =0.994, n =32, p <0.001) and negatively with 18:2n6 (LA) (r = − 0.921, n =32, p <0.001), and separated samples by month (June, July and August vs. September) (Fig. 8). PC2 correlated negatively with 16:0 (PA) (r = −0.799, n =32, p <0.001) and separated samples by position (inflow position vs. outflow position) (Fig. 9). Regarding lipid classes, the samples showed mainly high levels of pigments (chlorophyll and beta-carotene), glycolipids (mainly monogalactosyldiacylglycerol; MGDG and digalactosyldiacylglycerol; DGDG), phospholipids (mainly phosphatidylcholine; PC), flavonoids and sterols (Table 11). It is important to note that the pigment HPTLC layers appeared to overlap with those of other LC. The chlorophyll layer was overlapped with terpenes and beta-carotene with sterol esters, which affected the Table 7 Results of the two-way ANOVAs for nitrate, phosphate, nitrite and ammonium NUR.* Source of variation df SS MS F value P value Nitrate NUR System (S) 1 53.046 53.046 32.8891 0.000*** Time (t) 3 25.652 8.551 5.3015 0.0023** S ×t no significant Residual 74 119.353 1.613 Levene Test 7 3.8839 0.0012 Tukey Test DWC >NFT jun =jul =sept; jun <aug; jul =aug Phosphate NUR System (S) 1 0.3859 0.3859 11.735 0.0010** Time (t) 3 0.2377 0.0792 2.409 0.0744 S ×t no significant Residual 69 2.2694 0.0329 Levene Test 7 5.4699 0.000 Tukey Test DWC >NFT Nitrite NUR System (S) 1 0.0408 0.0408 18.8998 0.000*** Time (t) 3 0.0641 0.0214 9.8942 0.000*** S ×t 3 0.0225 0.0075 3.4757 0.0207 Residual 67 0.1447 0.0022 Levene Test 7 5.7592 0.000 Tukey Test DWC >NFT jun >jul; jun >aug; jun =sept; jul =aug =sept Ammonium NUR System (S) 1 0.0027 0.0027 8.5945 0.0046** Time (t) 3 0.0051 0.0017 5.4205 0.0021** S ×t no significant Residual 69 0.0215 0.0003 Levene Test 7 1.669 0.132 Tukey Test DWC >NFT jun >jul; jun =aug =sept; jul =aug; jul <sept NFT =nutrient film technique; DWC =deep water culture. * P <0.05. ** P <0.01. *** P <0.001. M. Castilla-Gavil´ an et al. Aquaculture 583 (2024) 740547 9 efficient quantification of these classes. Nevertheless, statistical analyses were carried out to determine whether the behaviour of these components in the samples actually corresponded to that of chlorophyll and beta-carotene. PERMANOVA showed differences in the lipid class profile between plants over the time (i.e. June, July, August and September; F =9.3306, p =0.001) and between plants in the two different positions (i.e. inflow and outflow; F =4.0790, p =0.008). The test also detected interactions between factors (F =3.3189, p =0.003), indicating that the differences between positions were present only in June, which is consistent with the results for total lipids (Fig. 7). PC1 of the PCA analysis explained 47.6% of the total variance and PC2 explained 15.51%. Phosphatidylcholine (r =0.840, n =32, p <0.001), MGDG (r =0.633, n =32, p < 0.001), and chlorophyll (r =0.815, n =32, p <0.001) correlated positively with PC1 and beta-carotene correlated negatively (r = −0.547, n =32, p <0.001), separating samples by time (Fig. 10, Table 11). Indeed, the presence of chlorophyll in the plants was significantly lower in September (two-way ANOVA, F =14.8046, p <0.001), while the presence of beta-carotene was significantly higher in this month (F =8.8260, p <0.001). In general, phospholipids and glycolipids followed the same trend. The presence of PC, PS, PI, PG +PE and MGDG increased over time until August and then decreased in September (Tables 11, 12). DGDG is an important LC with respect to MGDG and its presence remained mostly stable throughout the experiment (Tables 11, 12). Interestingly, the two-way ANOVAs for MGDG and flavonoids showed significant differences between plants in the different positions (F =7.3153, p =0.012 and F =9.4374, p =0.005, respectively; Table 11), these differences being present only in June for flavonoids (significant interaction between factors; F =10.0989, p <0.001). This is confirmed by the PCA, since MGDG correlates positively with PC1 (as explained above) and flavonoids with PC2 (r =0.327, n =32, p <0.05), tending to separate the samples according to their position in the system (Fig. 11). 3.6. Green leaf area Finally, to corroborate the results on the presence of pigments, the percentage of green leaves and stems relative to the whole plant was calculated for the DWC plants (Fig. 12). This percentage decreased significantly over time (Df =3, F =190.5772, p <0.001), but no differences were found between plants in the different positions in the system (inflow vs. outflow; Df =2, F =0.7606, p =0.4681). This is in agreement with the LC results, indicating that at least the chlorophyll/ terpenes and beta-carotene/sterol esters HPTLC layers showed a correct trend of pigment presence. 4. Discussion The findings of this study support the potential of integrating halophyte cultivation with fish RAS production in an IMTA system. In terms of plant productivity, the DWC system showed higher yields compared to the NFT system. This observation suggests that the DWC system Fig. 7. Total lipids content (%) of S. ramosissima edible leafs when cultivated in the DWC system during the trial. Error bars represent standard deviation. Inflow =plants placed in the tanks near to the water inflow; outflow =plants placed in the tanks near to the water outflow. Table 8 Results of the two-way ANOVA for total lipids.* Source of variation df SS MS F value P value Total lipids Position (P) 1 9.457e-06 9.457e-06 1.5936 0.2200 Time (t) 3 1.603e-04 5.343e-05 9.0046 0.000*** P ×t 3 9.227e-5 3.076e-05 5.1830 0.0073** Residual 22 1.306e-04 5.934e-06 Levene Test 7 0.3893 0.8986 Tukey Test jun: inflow>outflow jul, aug, sept: inflow =outflow Inflow =plants placed in the tanks near to the water inflow; outflow =plants placed in the tanks near to the water outflow. * P <0.05. ** P <0.01. *** P <0.001. M. Castilla-Gavil´ an et al. Aquaculture 583 (2024) 740547 16 Li, M., Callier, M.D., Blancheton, J.-P., Gal` es, A., Nahon, S., Triplet, S., Geoffroy, T., Menniti, C., Fouilland, E., Roque D’Orbcastel, E., 2019. Bioremediation of fishpond effluent and production of microalgae for an oyster farm in an innovative recirculating integrated multi-trophic aquaculture system. Aquaculture 504, 314–325. Lima, A.R., Castaneda-Loaiza, V., Salazar, M., Nunes, C., Quintas, C., Gama, F., Pestana, M., Correia, P.J., Santos, T., Varela, J., 2020. Influence of cultivation salinity in the nutritional composition, antioxidant capacity and microbial quality of Salicornia ramosissima commercially produced in soilless systems. Food Chem. 333, 127525. Lin, Y.-F., Jing, S.-R., Lee, D.-Y., 2003. The potential use of constructed wetlands in a recirculating aquaculture system for shrimp culture. Environ. Pollut. 123, 107–113. Lymbery, A.J., Doup´ e, R.G., Bennett, T., Starcevich, M.R., 2006. Efficacy of a subsurfaceflow wetland using the estuarine sedge Juncus kraussii to treat effluent from inland saline aquaculture. Aquac. Eng. 34, 1–7. Maciel, E., Lillebø, A., Domingues, P., da Costa, E., Calado, R., Domingues, M.R.M., 2018. Polar lipidome profiling of Salicornia ramosissima and Halimione portulacoides and the relevance of lipidomics for the valorization of halophytes. Phytochemistry 153, 94–101. https://doi.org/10.1016/j.phytochem.2018.05.015. Maciel, E., Domingues, P., Domingues, M.R.M., Calado, R., Lillebø, A., 2020. Halophyte plants from sustainable marine aquaponics are a valuable source of omega-3 polar lipids. Food Chem. 320, 126560. Marques, B., Maciel, E., Domingues, M.R., Calado, R., Lillebø, A.I., 2021. Halophyte plants cultured in aquaponics hold the same potential for valorization as wild conspecifics from donor sites. Appl. Sci. 11, 11586. https://doi.org/10.3390/ app112411586. Martins, C., Eding, E.H., Verdegem, M.C., Heinsbroek, L.T., Schneider, O., Blancheton, J.- P., d’Orbcastel, E.R., Verreth, J., 2010. New developments in recirculating aquaculture systems in Europe: a perspective on environmental sustainability. Aquac. Eng. 43, 83–93. Martins-Noguerol, R., Matías, L., P´ erez-Ramos, I.M., Moreira, X., Mu˜ noz-Vall´ es, S., Mancilla-Leyt´ on, J.M., Francisco, M., García-Gonz´ alez, A., DeAndr´ es-Gil, C., Martínez-Force, E., 2022. Differences in nutrient composition of sea fennel (Crithmum maritimum) grown in different habitats and optimally controlled growing conditions. J. Food Compos. Anal. 106, 104266. Mazliak, P., 1983. Plant membrane lipids : Changes and alterations during aging and senescence. In: Lieberman, M. (Ed.), Post-Harvest Physiology and Crop Preservation, Nato Advanced Study Institutes Series. Springer US, Boston, MA, pp. 123–140. https://doi.org/10.1007/978-1-4757-0094-7_5. Mesa-Marín, J., Mateos-Naranjo, E., Carreiras, J., Feij˜ ao, E., Duarte, B., Matos, A.R., Betti, M., Del Rio, C., Romero-Bernal, M., Montaner, J., Redondo-G´ omez, S., 2023. Interactive temperature and CO 2 rise, salinity, drought, and bacterial inoculation alter the content of fatty acids, total phenols, and oxalates in the edible halophyte Salicornia ramosissima. Plants 12, 1395. https://doi.org/10.3390/plants12061395. Meziane, T., Bodineau, L., Retiere, C., Thoumelin, G., 1997. The use of lipid markers to define sources of organic matter in sediment and food web of the intertidal saltmarsh-flat ecosystem of Mont-saint-Michel Bay, France. J. Sea Res. 38, 47–58. Mishra, A., Patel, M.K., Jha, B., 2015. Non-targeted metabolomics and scavenging activity of reactive oxygen species reveal the potential of Salicornia brachiata as a functional food. J. Funct. Foods 13, 21–31. Mizusawa, N., Wada, H., 2012. The role of lipids in photosystem II. Biochim. Biophys. Acta BBA-Bioenerg. 1817, 194–208. Mohammed, H.A., Emwas, A.-H., Khan, R.A., 2023. Salt-tolerant plants, halophytes, as renewable natural resources for cancer prevention and treatment: roles of phenolics and flavonoids in immunomodulation and suppression of oxidative stress towards cancer management. Int. J. Mol. Sci. 24, 5171. Murphy, J., Riley, J.P., 1962. A modified single solution method for the determination of phosphate in natural waters. Anal. Chim. Acta 27, 31–36. https://doi.org/10.1016/ S0003-2670(00)88444-5. Navarro, J.M., Flores, P., Garrido, C., Martinez, V., 2006. Changes in the contents of antioxidant compounds in pepper fruits at different ripening stages, as affected by salinity. Food Chem. 96, 66–73. Olsen, R.E., Henderson, R.J., 1989. The rapid analysis of neutral and polar marine lipids using double-development HPTLC and scanning densitometry. J. Exp. Mar. Biol. Ecol. 129, 189–197. Omoto, E., Iwasaki, Y., Miyake, H., Taniguchi, M., 2016. Salinity induces membrane structure and lipid changes in maize mesophyll and bundle sheath chloroplasts. Physiol. Plant. 157, 13–23. Padalino, L., Costa, C., Del Nobile, M.A., Conte, A., 2019. Extract of Salicornia europaea in fresh pasta to enhance phenolic compounds and antioxidant activity. Int. J. Food Sci. Technol. 54, 3051–3057. Pariza, M.W., Park, Y., Cook, M., 2000. Mechanisms of action of conjugated linoleic acid: evidence and speculation (44457). Proc. Soc. Exp. Biol. Med. 223, 8–13. Patel, M.K., Pandey, S., Brahmbhatt, H.R., Mishra, A., Jha, B., 2019. Lipid content and fatty acid profile of selected halophytic plants reveal a promising source of renewable energy. Biomass Bioenergy 124, 25–32. Rakocy, J.E., 2012. Aquaponics—integrating fish and plant culture. Aquac. Prod. Syst. 344–386. Ren, J., Han, E.J., Chung, S.H., 2007. In vivo and in vitro anti-inflammatory activities of α -linolenic acid isolated from Actinidia polygama fruits. Arch. Pharm. Res. 30, 708–714. Rodrigues, R.V., Schwarz, M.H., Delbos, B.C., 2007. Acute toxicity and sublethal effects of ammonia and nitrite for juvenile cobia Rachycentron canadum. Aquaculture 271, 553–557. Rozema, J., Flowers, T., 2008. Crops for a salinized world. Science 322, 1478–1480. Rozentsvet, O.A., Kotlova, E.R., Bogdanova, E.S., Nesterov, V.N., Senik, S.V., Shavarda, A.L., 2022. Balance of Δ5-and Δ7-sterols and stanols in halophytes in connection with salinity tolerance. Phytochemistry 198, 113156. https://doi.org/ 10.1016/j.phytochem.2022.113156. Sampaio, L., Wasielesky, W., Miranda-Filho, K.C., 2002. Effect of salinity on acute toxicity of ammonia and nitrite. Bull. Environ. Contam. Toxicol. 68, 668–674. Schindelin, J., Rueden, C.T., Hiner, M.C., Eliceiri, K.W., 2015. The ImageJ ecosystem: an open platform for biomedical image analysis. Mol. Reprod. Dev. 82, 518–529. https://doi.org/10.1002/mrd.22489. Schmautz, Z., Loeu, F., Liebisch, F., Graber, A., Mathis, A., Griessler Bulc, T., Junge, R., 2016. Tomato productivity and quality in aquaponics: comparison of three hydroponic methods. Water 8, 533. Schmautz, Z., Graber, A., Jaenicke, S., Goesmann, A., Junge, R., Smits, T.H., 2017. Microbial diversity in different compartments of an aquaponics system. Arch. Microbiol. 199, 613–620. Senff, P., Blanc, P.-P., Slater, M., Kunzmann, A., 2020. Low-technology recirculating aquaculture system integrating milkfish Chanos chanos, sea cucumber Holothuria scabra and sea purslane Sesuvium portulacastrum. Aquac. Environ. Interact. 12, 471–484. Shan, L., Oduor, A.M.O., Huang, W., Liu, Y., 2022. Nutrient enrichment promotes invasion success of alien plants via increased growth and suppression of chemical defenses. Ecol. Appl. n/a, e2791 https://doi.org/10.1002/eap.2791. Shi, C., Sun, G., Zhang, H., Xiao, B., Ze, B., Zhang, N., Wu, N., 2014. Effects of warming on chlorophyll degradation and carbohydrate accumulation of alpine herbaceous species during plant senescence on the Tibetan plateau. PLoS One 9, e107874. Shpigel, M., Ben-Ezra, D., Shauli, L., Sagi, M., Ventura, Y., Samocha, T., Lee, J., 2013. Constructed wetland with Salicornia as a biofilter for mariculture effluents. Aquaculture 412, 52–63. Sibiya, M., Sumbwanyambe, M., 2019. An algorithm for severity estimation of plant leaf diseases by the use of colour threshold image segmentation and fuzzy logic inference: a proposed algorithm to update a “leaf doctor” application. AgriEngineering 1, 205–219. https://doi.org/10.3390/agriengineering1020015. Sokal, R., Rohlf, F., 1995. Biometry, third. ed. WH Freeman, New York. Sousa, W.T.Z., Panitz, C.M.N., Thomaz, S.M., 2011. Performance of pilot-scale vertical flow constructed wetlands with and without the emergent macrophyte Spartina alterniflora treating mariculture effluent. Braz. Arch. Biol. Technol. 54, 405–413. Spiteller, G., 2003. The relationship between changes in the cell wall, lipid peroxidation, proliferation, senescence and cell death. Physiol. Plant. 119, 5–18. https://doi.org/ 10.1034/j.1399-3054.2003.00097.x. Sri-uam, P., Donnuea, S., Powtongsook, S., Pavasant, P., 2016. Integrated multi-trophic recirculating aquaculture system for nile tilapia (Oreochlomis niloticus). Sustainability 8, 592. Stankovi´ c, M.S., Petrovi´ c, M., Godjevac, D., Stevanovi´ c, Z.D., 2015. Screening inland halophytes from the Central Balkan for their antioxidant activity in relation to total phenolic compounds and flavonoids: are there any prospective medicinal plants? J. Arid Environ. 120, 26–32. https://doi.org/10.1016/j.jaridenv.2015.04.008. Surnar, S.R., Sharma, O.P., Saini, V.P., 2015. Aquaponics: innovative farming. Int. J. Fish. Aquat. Stud. 2, 261–263. Underwood, A., 1997. Experiments in Ecology: Their Logical Design and 977 Interpretation Using Analysis of Variance. Cambridge University Press, Cambridge. United Nations, 2023. Goal 12 | Ensure Sustainable Consumption and Production Patterns [WWW Document]. https://sdgs.un.org/goals/goal12 (accessed 10.17.23). Ventura, Y., Sagi, M., 2013. Halophyte crop cultivation: the case for Salicornia and Sarcocornia. Environ. Exp. Bot. 92, 144–153. https://doi.org/10.1016/j. envexpbot.2012.07.010. Sustainable cultivation and exploitation of halophyte crops in a salinizing world. Ventura, Y., Wuddineh, W.A., Ephrath, Y., Shpigel, M., Sagi, M., 2010. Molybdenum as an essential element for improving total yield in seawater-grown Salicornia europaea L. Sci. Hortic. 126, 395–401. https://doi.org/10.1016/j.scienta.2010.07.015. Ventura, Y., Wuddineh, W.A., Myrzabayeva, M., Alikulov, Z., Khozin-Goldberg, I., Shpigel, M., Samocha, T.M., Sagi, M., 2011a. Effect of seawater concentration on the productivity and nutritional value of annual Salicornia and perennial Sarcocornia halophytes as leafy vegetable crops. Sci. Hortic. 128, 189–196. Ventura, Y., Wuddineh, W.A., Shpigel, M., Samocha, T.M., Klim, B.C., Cohen, S., Shemer, Z., Santos, R., Sagi, M., 2011b. Effects of day length on flowering and yield production of Salicornia and Sarcocornia species. Sci. Hortic. 130, 510–516. https:// doi.org/10.1016/j.scienta.2011.08.008. Villette, C., Berna, A., Compagnon, V., Schaller, H., 2015. Plant sterol diversity in pollen from angiosperms. Lipids 50, 749–760. https://doi.org/10.1007/s11745-015-4008- x. Vizetto-Duarte, C., Figueiredo, F., Rodrigues, M.J., Polo, C., Reˇ sek, E., Cust´ odio, L., 2019. Sustainable valorization of halophytes from the mediterranean area: a comprehensive evaluation of their fatty acid profile and implications for human and animal nutrition. Sustainability 11, 2197. Waller, U., Buhmann, A.K., Ernst, A., Hanke, V., Kulakowski, A., Wecker, B., Orellana, J., Papenbrock, J., 2015. Integrated multi-trophic aquaculture in a zero-exchange recirculation aquaculture system for marine fish and hydroponic halophyte production. Aquac. Int. 23, 1473–1489. https://doi.org/10.1007/s10499-015-9898- 3. Walters, D., Raynor, L., Mitchell, A., Walker, R., Walker, K., 2004. Antifungal activities of four fatty acids against plant pathogenic fungi. Mycopathologia 157, 87–90. Wang, Y., Zhang, X., Huang, G., Feng, F., Liu, X., Guo, R., Gu, F., Zhong, X., Mei, X., 2020. Dynamic changes in membrane lipid composition of leaves of winter wheat seedlings in response to PEG-induced water stress. BMC Plant Biol. 20, 1–15. Webb, J.M., Quint˜ a, R., Papadimitriou, S., Norman, L., Rigby, M., Thomas, D.N., Le Vay, L., 2013. The effect of halophyte planting density on the efficiency of M. Castilla-Gavil´ an et al. Aquaculture 583 (2024) 740547 17 constructed wetlands for the treatment of wastewater from marine aquaculture. Ecol. Eng. 61, 145–153. https://doi.org/10.1016/j.ecoleng.2013.09.058. Wilson, C., Lesch, S.M., Grieve, C.M., 2000. Growth stage modulates salinity tolerance of New Zealand spinach (Tetragonia tetragonioides, pall.) and red orach (Atriplex hortensis L.). Ann. Bot. 85, 501–509. Zakynthinos, G., Varzakas, T., 2016. Carotenoids: from plants to food industry. Curr. Res. Nutr. Food Sci. 4, 38–51. Zhao, B., Wang, Li, Pang, S., Jia, Z., Wang, Lu, Li, W., Jin, B., 2020. UV-B promotes flavonoid synthesis in Ginkgo biloba leaves. Ind. Crop. Prod. 151, 112483. Zhongyan, F., Kefu, Z., 2003. Effects of photoperiod on flowering of Salicornia bigelovii Torr. Zhi Wu Sheng Li Yu Fen Zi Sheng Wu Xue Xue Bao 29, 474–478. M. Castilla-Gavil´ an et al.